German brewing and more - New pages [en]

PWM stir plate design

2013-05-07T02:45:55Z

Kaiser: /* part list */

{| style="width:800px"
|

=Introduction=

A stir plate uses a pair of spinning magnets to move a magnetic stir bar contained inside a flask or other vessel. The spinning of the stir bar keeps the liquid, in our case a yeast starter, agitated which promotes gas exchange and keeps the yeast in suspension. The main advantage of such a set-up is the ease with which it can be sanitized. All surfaces that come in contact with the starter, flask and stir bar, can be sanitized through boiling.

Experiments and observations by home brewers, including my own, have shown that constantly agitated starters increase yeast growth by 2 to 4 times over non agitated starters. This is the main reason why home brewers are interested in building or acquiring stir plates. Given the fairly high price ($50 for a simple design and $100+ for most commercial models) many home brewers op for building their own.

Commercial labs often favor the use of orbital shaker tables which also keep yeast cultures agitated but have the capacity to hold many flasks at once. Building a shaker table is more complicated than a stir plate which is why they are not commonly used by home brewers.

=Design options=

Most home built stir plates use a DC fan to which strong magnets are attached. The main design difference lies in how the speed of that fan is controlled. Simple designs put a variable resistor between the DC power supply and the fan. Slightly more complex designs use a linear voltage controller to regulate the voltage applied to the fan. The problem with designs that control fan speed through voltage is that most fans have a narrow voltage band between no rotation and full speed. This makes speed control difficult unless there is sufficient resistance from the liquid that is stirred.

This speed control problem is overcome by pulse width modulation (PMW), which does not change the voltage applied to the fan but the amount of time the fan is turned on and off. The frequency of the pulsed fan power is generally between 10-100 Hz. A PWM based design is described by this article.

An even more sophisticated stir plate control uses a micro controller and speed sensor to implement a feedback loop that allows for accurate fan speed control based on a speed (RPM) set by the user. The Digital Stirplate offered by [http://www.digitalhomebrew.com/p/52/digital-stirplate Digital Homebrew] employs such a design.

=PWM design=

[[File:555 internals.gif|frame|right|Figure 1 - the internals of the NE555N timer chip connected to a capacitor that can be charged and discharged through a variable resistor]]

The fan speed control logic described here is based on a [http://www.homebrewtalk.com/f51/simple-pwm-stirplate-controller-219121/ Home Brew Talk post] by rocketman768 with a few modifications. It employs a [http://www.mouser.com/ds/2/389/CD00000479-103226.pdf NE555N] timer chip. The internals of the NE555N are basically a RS flip flop with differential comparators its R (reset) and S (set) inputs as shown in figure 1. When the voltage at the THRES input exceeds the internally generated Vthres the R input of the flip flop asserts high and the output is reset to 0. Conversely if the voltage at the TRIG input is lower than the internally created Vtrig the flip flop's S input asserts high and the output Q is asserted high. In this design TRIG and THRES are both connected to the same terminal of a capacitor and as a reslt R and S can never be asserted at the same time.

Pulses of differing witdth will be created by triggering the R and S inputs at changing time intervals through charging and discharging of the capcitor. The speed which which a capacitor charges depends on the product of its capacitance and the resistor though which it is charged, also called RC constant. Changing the capacitance of a capacitor is difficult but changing the resistance of a resistor is much easier, which why the capacitance remains constant but the resistance is changed.

[[File:PWM detail.gif|frame|center|Figure 2 - Charging and discharging the capacitor causes the generation of pulses on the Q output. The width of these pulses depend on R1 and R2 which are based the current position of the potentiometer]]

In this design the capacitor is connected to the output Q though variable paths of a potentiometer and diodes. Figure 2 illustrates how the capacitor is charged and discharged as a result of being connected to Q. When Q is asserted high, the cpacitor is charged through the R1 part of the potentiometer until the voltage on the capacitor reaches Vthres at which point the R input of the flop is triggered and Q asserts low. Now the capacitor discharges through the R2 part of the potentiometer until the capacitor voltage falls below Vtrig, S is triggered and Q asserts high repeating the process of charging the capacitor.

The result is that the time of Q being asserted high deepends on the product of R1 and the capacitance C while the width of Q being asserted low depends on R2 and the capacitance C. Since R1 and R2 are part of a potentiometer they are adjustable but their sum has to remain constant. What follows is that changing R1 changes the percentage that Q is asserted high but not the frequency of pulses on Q.

To control the fan, Q is used to drive a MOSFET which turns the fan on and off. The longer Q is asserted high the longer the fan is turned on and thus the faster the fan will spin. The interia of the fan coulped with a sufficiently high pulse freqiency resuts in an even rotation speed despite the pulsed nature of fan's power supply.

==Wiring diagram and parts list==

[[File:StirPlatePWM wiring.gif|frame|center|Figure 3 - Wiring diagram of the PWM control logic and the DC fan]]

Figure 3 shows the complete schematic for the control logic and the fan. 555 is the timer chip. D1 and D2 control which section of the potentiometer P1 controls charging and discharging of C1, respectively. D3 is a diode that protects the MOSFET from voltage spikes that happen when the current through the inductive load, the fan, is suddenly interrupted. Capacitors C2 and C3 stabilize the input power supply.

[[File:StirPlatePWM breadboard.gif|frame|center|Figure 4 - suggested breadboard layout]]

Figure 4 shows the suggested layout on a breadboard. A standard breaboard is large enough for 2 control circuits which is why the option for a 2nd control circuit is shown. Pins 2 and 6 of the 555 are connected with a piece of wire under the breadboard (red dashed line) while all other connections are made with wire jumpers on top of the board (blue lines). The layout also shows the connections within the bread board for reference (thin gray lines). (+) and (-) are the 12 V power supply connections, M+ and M- are the fan terminals and M,L,R refer to the middle, left and right terminals of the potentiometer. If the fan speeds up when the potentiometer is turned to the right, simply reverse the L and R connections.

===part list===

The following is a part list including [http://mouser.com mouser.com] item numbers. The part numbers are given for reference and no guarantee is made for their correctness.

{| class="wikitable"
|-
! id !! description !! mouser.com model item number
|-
| 555 || NE555N timer chip || 511-NE555N
|-
| D1, D2, D3 || general purpose diode || 625-1N4933-E3
|-
| C1 || Aluminum Electrolytic Capacitors, 2.2 uF || 647-UVY2A2R2MDD
|-
| P1 || 100k linear potentiometer || 858-P160KNP0C20B100K NOTE: I was not able to find a good knob for this potentiometer.
|-
| C2 || Aluminum Electrolytic Capacitors, 22 uF || 140-RGA220M2ABK0811G
|-
| C3 || Multilayer Ceramic Capacitor, 0.1 uF || 810-FK18X7R1E104K
|-
| Q1 || MOSFET NFET DPAK 30V 54A 5.5 mOhm || 863-NTD4906N-35G
|-
| M || 12 V DC fan || 670-OD8025-12HSS
|}

You'll also need a 12V DC power supply and a jack for connecting this power supply.

==Housing==

Home brewers have been very creative when it comes to mounting the fan in a housing. As long as the top cover, which will be between the fan and the flask, any box will do. While some have even used Tupperware (R) containters they do not necissarily provide enough rigidy to support a 2 L or lager Erlenmeyer flask. I suggest using a sturdy pastic project box or custom build enclosure. The double stir plate shown here uses a simple housing fashioned from thick plywood and acryllic glass. Being able to see the spinning fan can be an advantage in some cases.

[[File:StirPlatePWM-1.jpg|frame|center|Picture of the completed double stir plate. This design is 14 inches wide, 7 inches deep and 2.5 inches high. ]]

[[File:StirPlatePWM-2.jpg|frame|center]]

==Testing & troubleshooting==

The more complicated the control the more chances there are for defects. While an oscilloscope would be very useful in debugging this circuit, a simple voltmeter will do as well.

Once completed apply the 12 V supply voltage. If nothing starts smoking see if turning the potentiometer changes the fan speed. If nothing happens check the supply voltage on the 555 (pin 8) and the reset pin (pin 4). They should both be at 12 V. The check that GND (pin 1) is at 0 volt. Now pin 3. If that's at 0 V the fan should be off and when it is at 12 V the fan should be running at full speed. Now check that both pins 6 and 2 have the same voltage as pin 3. If they don't check their connections through the potentiometer.

|}

Lactate Taste Threshold experiment

2013-03-10T13:50:37Z

Kaiser:

{| style="width:800px"
|

Lactic acid is a very popular acid for brewing water treatment and mash pH correction. Home brewers are using it in the form of acidulated malt or 88% concentrated lactic acid. A common concern with lactic acid is its impact on the final beer flavor when rather large amounts are used. Lactic acid is associated with sour beers and that flavor may not be welcome in a Pilsner style beer, for example.

The question many brewers are struggling with is: How much lactic acid is needed to make a negative impact on the flavor of the beer while the mash pH is still in the desired range ?

The literature provides little information on this topic. "Brewing: Science and Practice" lists a lactic acid taste threshold of 400 mg/l<ref name="Briggs">Dennis E. Briggs, Chris A. Boulton, Peter A. Brookes, Roger Stevens, ''Brewing Science and Practice'', Published by Woodhead Publishing, 2004</ref>. No detail is given about the experiments. It's unknown if during these experiments the lactic acid was neutralized. This aspect is important since the sour taste of lactic acid comes from its ability to lower pH. When neutralized to the pH of the sample in which it is tasted lactic acid is more difficult to detect.

Narziss and Back mention acidulated malt additions at levels as high as 8% to neutralize water alkalinity and lower the mash pH to a desirable level of 5.5 - 5.6<ref name="Narziss_Back">Ludwig Narziss, Werner Back, Die Bierbrauerei Band 2: Technologie der Würzebereitung, Wiley-VCH, 2009</ref>. For a 12 Plato beer brewed with 95% efficiency into kettle this adds 380 mg/l lactate to the beer.

A common upper bound of acid malt use given by home brewers is 4-5%.

To answer the question of taste threshold of lactate in water and beer I conducted a taste experiment with 8 members of my home brew club [http://www.bfd.org Brew Free or Die] as panelists. Each panelists received 4 sets of 6 different samples and was asked to sort them into one group in which the off flavor is not detectable and sort the rest by the intensity of that flavor. The panelists were not told what the off flavor is but knew which sample was the control.

= Materials and Methods=

The four sets of samples were water, Bud Light, Budweiser and Sierra Nevada Torpedo Ale. I added increasing amounts of calcium lactate to these beers and water. It was important for this experiment that the addition of lactate did not change the pH of the sample which is why I prepared a 23 g/l lactic acid solution by adding 8.25 ml 88% lactic acid to 375 ml reverse osmosis water. A sample of that solution was taken and dry slaked lime (Ca(OH)2) was added until it reached the pH of the repective water or beer sample. If the pH was too high a bit more lactic acid solution was added to bring it down again. By doing so the lactate/lactic acid concentration did not change.

The resulting solution was then added to the samples. For the beers a mix of water and Ca-lactate solution was added to keep the total addition of water the same.

Water

{| class="wikitable"
|-
! sample ID !! volume (ml) !! water added (ml) !! Ca-lactate added (ml) !! resulting lactate concetration (mg/l)
|-
| 0 || 600 || 0 || 0 || 0
|-
| 1 || 600 || 0 || 5 || 193 (labeled 200)
|-
| 2 || 600 || 0 || 10 || 387 (labeled 400)
|-
| 3 || 600 || 0 || 18 || 696 (labeled 700)
|-
| 4 || 600 || 0 || 25 || 968 (labeled 1000)
|-
| 5 || 600 || 0 || 35 || 1239 (labeled 1200)
|}

Bud Light, Budweiser and Sierra Nevada Torpedo

{| class="wikitable"
|-
! sample ID !! volume (ml) !! water added (ml) !! Ca-lactate added (ml) !! resulting lactate concentration (mg/l) !! equivalent amount of acidulated malt for 12 Plato beer*
|-
| 0 || 350 || 18.7 || 0 || 0 || 0%
|-
| 1 || 350 || 15.8 || 2.9 || 193 (labeled 200) || 3.7%
|-
| 2 || 350 || 12.8 || 5.8 || 387 (labeled 400) || 7.3%
|-
| 3 || 350 || 8.2 || 10.5 || 696 (labeled 700) || 13.2%
|-
| 4 || 350 || 4.1 || 14.6 || 968 (labeled 1000) || 18.3%
|-
| 5 || 350 || 0 || 18.7 || 1239 (labeled 1200) || 23.4%
|}

<nowiki>*</nowiki> the equivalent amount of acidulated malt assumes an efficiency into the kettle of 85% and a lactic acid content of 3% (w/w) for the acidulated malt.

Samples were assigned random letters with the exception of 0, which always received letter A. Each panelist received a score sheet on which he/she would note which samples taste like A and the intensity ranking of the off-flavor in the other samples.

Samples were presented in this order: water, Bud Light, Budweiser, Sierra Nevada Torpedo Ale.

Some of the panelists were BJCP certified judges.

= Results and Discussion =

One theme apparent during tasting was how much difficulty panelists had in consistently identifying the samples with the most added lactate. This provided a challenge in deriving trends from the taste results.

Figure 1 shows the tasting results for the 4 sets. An intensity of 0 was assigned to all samples that a given panelist indicated as tasting like the control. The sample with the most intense flavor was assigned a 5, the second most intense a 4 and so forth. The problem with this approach is that many panelists did not identify the sample with the most lactate as the one with the most intense off flavor. Panelist 6, for example noted that the water with 200 ppm lactate tasted most intense.

To account for this I took the lactate level of the sample that was noted as having the most intense "off-flavor" and used it as the level for the most intense tasting sample indicated by that taster. The intensities of the other samples were then scaled based on that. That resulted in the charts in Figure 2.

[[File:LactateTaste_charts_1.gif|frame|center|Figure 1 - The off-flavor rating given to the different samples by the 8 panelists. Samples reported as tasting like the control are shown as 0 while the sample with the most noticeable off-flavor is shown as 5. See text for further detail]]

[[File:LactateTaste_charts_2.gif|frame|center|Figure 2 - scaled flavor intensity was determined by using the lactate level of the sample that a given panelist reported as tasting the most intense as the upper limit for that panelist.]]

Just by looking at the density and height of bars, lactate is more easily detected in water. This does not come as a surprise. Some identified it down to a level of 200 mg/l, while some where not able to detect it until it reached levels of 700 mg/l.

Panelists had more difficulty identifying the flavor in the 3 beers that were presented. To my surprise Bud Light did the best job of hiding this flavor. This may have been the result of panelists being unfamiliar with the taste of this beer, something that should also be true for Budweiser, or that it was the first flight of beer they were presented with.

Another explanation is that panelists had general difficulties identifying the flavors in the beers even when the highest amount added was 1200 mg/l. This becomes evident by the rather even spread of high ratings (5) over the 200-1200 mg/l range in Figure 1. In water those higher bars are more clustered toward the higher lactate range.

Based on this and having tasted lactate in Budweiser myself, I'm willing to conclude that it is fairly difficult to detect added lactate at a level of 400 or below.

This experiment was not able to show that the intensity of the beer flavor changes the perception threshold for additional lactate, although it showed that it is more easily detectable in water than in beer. It seems logical that more intensive beer flavors would be able to mask higher levels of lactate, however, more experimentation is needed for this, in particular the maximum lactate level should be raised to allow it to stand out more prominently as many panelists misidentified the sample with the most lactate added.

Figure 3 is yet another way to look at the results. For each series and panelist it shows the highest level of lactate that was marked as tasting like the control. The correlations is not very strong but it shows that lactate was more easily detected in water. The high bars for Bud Light show that panelists had trouble detecting lactate in that beer. While generally low bars for Budweiser and Sierra Nevada Torpedo suggest that it was easier to detect in these beers. However, this is likely a random error since the low correlation between actual lactate levels and flavor rating suggests that most panelists ended up guessing or mistook other beer flavors for the off flavor. Palate fatigue is also an issue in this experiment due to the fairly large number of samples that were tasted.

[[File:LactateTaste_charts_3.gif|frame|center|Figure 3 - The highest level of added lactate that was a given panelists reported as tasting like the control. The higher the bars the more difficult it was to detect the lactate]]

= Conclusion =

It was surprisingly difficult for panelists to pick out beers that had lactate added even at levels that correspond to an equivalent acidulated malt use of 13% and higher. Note that the acidity of the lactic acid was neutralized with slaked lime. A general recommendation for home brewers is to keep the use of acidulated malt below 5%, which corresponds to a level of 264 mg/l added lactate in a 12 Plato beer with 85% efficiency into kettle. Many of the panelists were not able to pick up the added lactate at a level of about 400 mg/l which corresponds to about 7.5% acidulated malt. Based on that we can safely say that even 8% acidulated malt won't ruin a beer if that amount is needed to counteract water alkalinity.

= References =

<references/>

|}